Third World Tourism

I got a pineapple for 10 pesos, or 40 cents USD. Bananas are 2 cents each.

I just returned from a country where the internet consists of tin cans connected with bits of string. It is a nation trapped in 1960, the last year it could legally import goods from the free world.

The city air is choked with diesel exhaust from six-bangers built in the height of last century. The resident nationals don’t know that they are frozen in time. Their only window to the outside world comes from five government-curated television stations.

The national economy is meager by Western standards, but the citizens are comfortable. The government meets basic needs. Provisions are sold at ration shops at government-mandated prices. Rice is 0.70 pesos per pound, which converts to 3 cents USD. A dozen eggs is 2 pesos, or 8 cents USD. The first six eggs each month are free.

Ration shop. It's like Whole Foods, but cheaper. — Ration shop. It’s like Whole Foods, but cheaper.

Everything is cheap, including wages. The average income is $25 USD per month. They don’t know how much Americans make. The only way to feel poor is to expose yourself to the rich.

Occasionally, they are afforded a glimpse of the first world when overseas visitors pass through their lives. Those who have encountered tourism don’t quite understand it, but want it. Locals have learned that outsiders invariably come with fat wads of cash. They know that in the next year or so, the US ban on tourism will be lifted, and expect floods of first-world money to swarm into their humble universe.

They’re getting ready for this. Women are putting signs on their front doors indicating that the home is a casa particular – a.k.a. bed & breakfast. Bus operators prohibited from transporting non-nationals have learned to drop tourists off just outside city limits. They pocket the higher cash fare.

This sign indicates a casa particular. Visitors can knock on the door and ask for a room for the night. — This sign indicates a *casa particular*. Visitors can knock on the door and ask for a room for the night.

The problem with third-world tourism is that it creates a mining economy with the sole purpose of extracting money from visitors. There is a pronounced wealth inequality between those who receive money from tourists and those who don’t. It turns women into prostitutes and men into thieves. The main exports are counterfeit goods and STDs. See also: Bangkok. Hanoi. Tijuana.

At best, the end product is a culture devolved into a sea of obsequious pressed uniforms. Then the only homes are hotels and the only industry is hospitality. The main exports are souvenir mugs and inflated self-importance. See also: Singapore. Bali. Cancun. The biggest import is cultural imperialism and the domestic product is resentment.

Maybe isolation isn’t so bad after all.

Home sweet home

A residential street in El Centro

Cafeteria Menu: 1 peso for a coffee, 5 pesos for a ham sandwich

Pizza shop with 40 cent pizzas

1 peso (4 cents USD) for an ice cream cone

January 16, 2015

Baby Startups: Handle with Care

I don’t know how humans haven’t driven themselves to extinction.

Breeding is easy, my unwed cousin can tell you that. But the resulting offspring have no survival skills whatsoever. A baby horse can run with the herd two hours after birth. A newborn human, left to its own devices, will crap itself and die.

Today was our final day at AOL’s First Floor Labs. As we were saying our goodbyes, I realized that everything had changed. No company was leaving with the same team they had six months ago. Some had vaporized completely. I’m not one to talk, because we are no longer the company we arrived with.

At my age, 6 months feels like nothing. For a startup, 6 months is a death-defying feat.

Paul Graham says that inexperienced founders and investors forget just how fragile startups are. We spend our days scheming ways to take over the world and forget basic care and feeding instructions in the process.

Startups are newborn babies, they need to be kept in incubators with adult supervision. We’ve moved out of our First Floor Labs office, but we’re going someplace bigger and better. In spite of our best efforts, we’re not dead yet.

January 10, 2015

How Bitcoin Mining Makes the World a Better Place

To mine a single bitcoin generates the same carbon footprint as burning 15.9 gallons of gasoline, and Bitcoin mining consumes $15 million worth of electricity every day. People have criticized Bitcoin for not being very green. We have massive data centers around the world with massive computational power, and they’re doing nothing but non-stop hash functions.

On September 11, 2001, two towers full of people and data collapsed to the ground. The New York Stock Exchange, NASDAQ, NY Mercantile Exchange, and US bond markets all stopped trading for four days.

They had to close because the passwords to the nation’s wealth accounts vaporized with the managers who were working in the World Trade Center. Without access to the market-makers’ accounts, the exchanges couldn’t open for trading.

The only way to recover a password is through brute force hash functions. This was a lot to ask of 2001 computational technology.

To save time, security experts called up distressed families to ask them for kids’ names, pets’ names, ex-girlfriends, childhood best friends, anything that could potentially be used as part of a password. It was only with huge amounts of help that passwords were recovered in time to reopen markets the following week.

Today, none of that would be necessary. Thanks to advancements in Bitcoin mining technology, you can rent time on a Bitcoin cloud and hash out 10-character passwords in just a few hours.

We no longer need to worry about whether our families can access our accounts in event of an untimely death. New technology often appears deceptively useless.

See Also:
The magic of mining –Economist

January 9, 2015

Mining Bitcoin and Cracking Passwords

Did you think people were buying high-power wifi adapters for the better signal? — Did you think people were buying these for the signal strength?

I spent the last few evenings mining passwords from the wifi networks in my new neighborhood. I was just testing them, see, because I’m really concerned for my neighbors’ security and such.

It used to be that WEP was the only form of wifi security, but it turned out to be too easy to recover a WEP key: Just eavesdrop on the network and examine the frequency of repeated bytes in captured packets.

Now all routers use WPA, which hashes the password using the SHA1 function.

It’s still easy to sniff a packet and get the encrypted key, but the only way to pull the password from the key is through brute force: Run the SHA1 function on combinations of letters until you get something that results in the same key.

At this point you might be thinking, Geez Elaine, just pay for freaking Comcast already.

Brute-forcing a password might sound like a huge waste of computational resources, but this is exactly what Bitcoin miners do. Mining is the process of guessing a password called the nonce. The nonce is a number that can be appended to a block and hashed, resulting in something that starts with a string of 0s. If a miner finds a nonce, they win the block and get 25 bitcoins.

So I spun up a GPU instance on AWS. This was the same instance I had used to mine Bitcoin before Amazon shut it down.

More than half the time, the wifi password is a phone number. Hashing all combinations of 10-digit numbers is quick. The rest of the time, the password contains proper nouns. Even with GPU acceleration, hashing every possible combination of characters takes days.

And that is why I didn’t have internet access until today.

December 25, 2014July 23, 2015

Deep Learning vs. Hierarchical Models for M&M Identification

My parents keep a candy jar on the coffee table. It usually contains M&M’s.

This past week the jar had almond-looking M&M’s in Christmas colors, and I stuffed a handful into my face because that’s how M&M’s are meant to be eaten.

I did not notice that the M&M’s were missing the “m”.

Turns out they weren’t almond M&M’s. They were JORDAN ALMONDS, which are a heathen abomination.

This is troubling, because I am a Bayesian learning machine. If you were to dissect my brain and model it as a Bayesian network, it would look like this:

Pastel + oblong-shape is almost certainly a Jordan almond, which we need to avoid.

Bayesian learning updates a hypothesis in terms of how well it can explain observed data. Each candy I eat reinforces or weakens the dependency between observations, depending on whether or not it is a flavor I like.

The first observation is color. The color scheme affects each observable candy shape differently. Standard M&M colors (red-orange-yellow-green-blue-brown) applied to an oblong-shaped candy has always been observed to be an almond M&M, which are dericious. However, standard M&M colors applied to fat-shaped candy means that it is likely a peanut butter or pretzel M&M, both of which are terrible. An observation of green/white or green/brown colors on fat-shaped candy indicates mint or coconut M&M’s. Those are awesome.

From top left: Peanut = NO; Coconut = awesome; Pretzel = NO; Mint = YES x100; Peanut butter = gnarsty; Dark choco peanut = NOO; Candy corn = cat vomit; Almond = YESSSSS; Dark choco = Yes; Normal = Yes.

Color observation has weak interaction with round-shaped candies, because round M&M’s can be printed in custom colors. Thus the round shape requires an “m” label observation. Round candies without a label might be M&M knockoffs, or far worse: Reese’s Pieces.

M&M knockoffs, but probably better than Reese's Pieces — M&M knockoffs, but probably better than Reese’s Pieces

I ran into trouble here when I observed seasonal colors on an oblong shape, which, according to 100% of prior observations, should have been holiday almond M&M’s. If I update my model to reject all unlabeled oblong candies, I would also lose out on plain chocolate-covered almonds.

The problem lies in the fact that in my Bayesian model, there is no direct interaction between the first and third layer, color and “m” label.

Unlike Bayesian hierarchical models, Deep Learning machines do not start by assuming the structure of interactions between observations. Maybe classification by color and shape is not the best way to represent my data. Maybe I should combine observations in different ways.

An untrained deep learning machine looks something like this:

Within the octagon are hidden layers that factor all possible polynomial interactions between observations. As the machine is trained, irrelevant interactions disappear from the model and relevant interactions develop stronger connections.

Eventually, each layer becomes a different representation of the observations. This is similar to how a block of text can be represented as groups of characters in one layer, words in another, or sentences in a third. I don’t know how my brain will ultimately represent M&M nodes. Maybe some third-order function of color, size, and time of year, or maybe something far more complex.

A Deep Learning machine requires far more data to train, which means I need to eat a lot more candy if I want to do some Deep Learning.

Elaine's Idle Mind

and Devil's Workshop

Third World Tourism

Like this:

Baby Startups: Handle with Care

Like this:

How Bitcoin Mining Makes the World a Better Place

Like this:

Mining Bitcoin and Cracking Passwords

Like this:

Deep Learning vs. Hierarchical Models for M&M Identification

Like this:

Go talk about it:

Like this:

Go talk about it:

Like this:

Go talk about it:

Like this:

Go talk about it:

Like this:

Go talk about it:

Like this: